Insights by Roy Schoenberg, M.D., CEO, Aileen, and Founder and Executive Director, Amwell
Key Takeaways
- Telehealth failed because it replicated visits instead of redistributing expertise.
- AI adoption will be driven by cost incentives.
- Payers may launch AI-first insurance plans.
- Seniors need relationship-based AI, not reminder apps.
- AI will become healthcare’s “surround system.”
“Telehealth can be a channel, or it can be a switchboard.”
That’s how Dr. Roy Schoenberg frames one of the most important misunderstandings of virtual care. For more than a decade, telehealth has largely been implemented as a substitute for the traditional office visit, basically, the same encounter, delivered through a screen. Useful, convenient, and often necessary, but ultimately limited.
Schoenberg argues that this isn’t what telehealth was supposed to become.
A physician by training and one of the pioneers who brought telehealth into mainstream U.S. healthcare through Amwell, he has spent his career operating at the intersection of clinical care, technology, policy, and large-scale systems. After leading Amwell through its growth into a foundational platform for payers and health systems, he stepped into his next venture with a bold thesis. AI will become the primary entry point to healthcare, and its biggest early impact may come not from flashy interfaces, but from simple, relationship-driven interactions, especially for seniors.
His new venture, Aileen.ai, is built around that belief. And the story he told our host on a recent episode of the Big Unlock Podcast isn’t a “telehealth recap.” It’s a provocative forecast of how AI will reshape the economics of access, the structure of care journeys, and the missing human support that millions of seniors increasingly need.
Listen to the full conversation
Telehealth’s missed opportunity: replacing visits instead of redistributing expertise
Schoenberg’s answer to host Ritu M. Uberoy’s first question is direct: Telehealth has largely been used as another channel for the same clinical encounter. A visit or follow-up visit that used to happen in the office now happens via video. That creates convenience and some efficiency. But it doesn’t fundamentally change the healthcare system. He refers to this as “model one.”
He then presents his “model two,” which he says is where the disruption lives. Telehealth “as a switchboard’ that reshuffles how expertise is acquired and delivered.
He offers a clear example. If you live in Boston, cancer expertise is concentrated in world-class centers. But many parts of the country don’t have that depth of specialty care. Telehealth’s deeper promise is not just letting a local doctor do a video visit; it’s enabling expertise to flow at scale, so patients and clinicians in underserved regions can access high-quality specialty care without relocating.
In his words, this would “democratize” services at a much larger scale.
So why hasn’t it happened?
Schoenberg points to the system’s protectiveness and the “muscle memory” of healthcare. Licensure, credentialing, reimbursement, and entrenched workflows make it difficult to redistribute expertise. He’s candid about underappreciating how resistant the system would be to this kind of restructuring.
But he also makes a strong claim: the train has left the station. Post-COVID, the idea that care will be redistributed through technology extends beyond any one platform or company. It may move slowly, but it’s inevitable.
That sets up his next point: AI will accelerate the switchboard model far more than traditional telehealth ever could.
AI will become the front door because of economics, not because it “beats doctors”
A lot of the public debate about healthcare AI gets stuck in a single question: can AI be a doctor?
Schoenberg doesn’t dismiss that question, but he argues it’s not the real trigger for adoption.
He believes AI will become the primary entry point to healthcare for a simpler reason: economics. AI is highly accessible and far cheaper than clinician time. That means the earliest large-scale adoption will be driven less by persuasion —“AI is better than your doctor”—and more by incentives—using AI saves you money”.
His most provocative prediction is that the true inflection point will come when payers introduce an insurance product that requires members to interact with AI first. He compares it to the gatekeeper model of the HMO era, except the gatekeeper won’t be a primary care physician. It will be AI.
He acknowledges that pure restriction won’t be popular. The product will need to be designed “smartly,” with a break-glass option to see a clinician when needed. But the direction is clear: the more you use AI, the more you save. Shared savings models and cost-driven pathways will shape behavior.
This matters because it reframes “AI adoption” as a system design and insurance design problem ,not just a clinical intelligence problem. If the payer controls the front door and aligns incentives, AI doesn’t have to win a philosophical argument. It wins by being the default.
Schoenberg also notes a reality we’re already seeing. People are using general AI tools for health questions at scale, often as a preparatory step before seeing a clinician. That creates a gradual normalization effect. Patients build comfort by asking questions, receiving explanations, and forming a first draft of what they want to discuss.
In his view, this is the beginning of AI as the “surround system” for the healthcare experience, an inevitable layer that wraps around access, triage, navigation, and follow-up.
Why “staying power” is the real breakthrough for seniors and caregiving
If the first half of the episode is about how AI will restructure the front door, the second half is about a different problem entirely. The “caregiving gap.”
Schoenberg describes a sobering demographic reality. The senior population is growing rapidly, while the availability of caregivers is shrinking. Senior care is emotionally and physically demanding and often underpaid. Many people don’t want those jobs. The gap between need and available support is widening.
His claim is blunt. “Houston, we have a problem…”
There’s already a massive “age tech” market, apps, chatbots, talking devices, pill boxes, and reminder tools. Many are well-intended, but he argues most fail for one consistent reason. They don’t create adoption or “staying power.” Seniors have a complicated relationship with technology, and tools that feel like nagging reminders create fatigue. They “die on the vine.”
His latest venture, Aileen.ai, is built around a different premise.
If technology is going to meaningfully influence a senior’s life, the first job isn’t telling them what to do. The first job is becoming a wanted presence in their day, something they choose to engage with.
Schoenberg defines staying power as something that comes from familiarity and relationship. In real life, staying power comes from people who know you, who remember your kids and grandkids, your joys, frustrations, and stories. Aileen is designed to create that kind of familiarity through AI.
And he emphasizes the real technical challenge: none of that “personal narrative” exists in a database. Nobody wrote a book about your dad. So if AI is going to know a senior deeply, it has to learn that reality over time.
He also points out a behavioral constraint. Most AI today is prompt-driven. We type something in, and AI responds. That interaction model won’t work for seniors. If you wait for seniors to prompt, you’ll wait forever. So Aileen is designed to initiate engagement.
That’s why the interface choice matters. Aileen uses the phone. It calls seniors. It doesn’t require them to download an app, log in, pair devices, or even have Wi-Fi. The “backend” may be rocket-science AI, but the front end is intentionally simple and familiar.
Schoenberg calls this combination “schizophrenic” in the best way. Delivering “science-fiction technology” behind a human, everyday interface.
He also describes another distinctive element. Aileen builds intimacy by learning from the people who already know the senior. It can call family members casually, without forms or scheduled meetings, to gather context and build an understanding of the senior’s life. Only after it crosses a threshold of “knowing enough” does it begin daily engagement with the senior. Then it loops back insights to family members. He describes it as like having a lightweight companion and monitoring layer that helps shoulder the burden families carry.
Critically, Aileen isn’t designed to talk about “healthcare” all day. It’s designed to talk about what seniors want to talk about, because relationships are what create engagement. Once that staying power exists, Aileen becomes a mouthpiece for other healthcare technologies: reminders, symptom monitoring, mood and cognitive signals, and supportive guidance.
Schoenberg’s bet is that relationship is the missing prerequisite to successful senior-facing health technology. Without it, the reminders don’t stick. With it, they do.
AI is inevitable, but it will mature through trial, error, and redesign
Schoenberg closes with a realistic forecast. AI is young. We will see a long maturation curve. There will be mistakes. There will be things to worry about. But he believes AI’s role as a foundational “surround system” in healthcare is inevitable.
His message is basically: if we know we’re going there, we have to start walking.
That’s what he believes Aileen represents: an early attempt to solve a hard problem the system can’t ignore: a widening caregiving gap and a need for technology that doesn’t just function, but persists in daily life.
He’s confident in the ambition: in his words, this could change the world “no less than what telehealth did.” Whether one agrees with the magnitude or not, the through-line is consistent: the next era of healthcare won’t be defined by a single app or a single visit channel. It will be defined by AI as the first touchpoint, the navigation layer, and perhaps for the most vulnerable among us, an ongoing relationship that helps people stay supported at home.
The Takeaway
Dr. Roy Schoenberg’s message is both pragmatic and bold. Telehealth’s real promise was never just “video visits.” It was the ability to redistribute expertise and reshape care journeys at scale, and AI will finally push healthcare toward that switchboard model by changing the economics of access. In his view, AI won’t become dominant by proving it is “better than doctors,” but by becoming the default entry point through payer-driven incentives that reward AI-first navigation while keeping a break-glass path to clinicians.
Sitting at the intersection of telehealth platform-building and the next wave of AI-driven care navigation and companionship, Dr. Schoenberg’s unique insights are especially valuable:
- Telehealth was mostly used as a substitute channel; its deeper potential lies in acting as a switchboard that redistributes expertise and democratizes access.
- AI will become healthcare’s front door primarily because of economics—accessibility and cost—not because it “proves” it’s better than doctors.
- The real adoption trigger will be payer products that require AI-first interaction, with shared-savings incentives and a “break-glass” path to clinicians.
- The senior care crisis is a demographic reality: need is rising while caregiver supply is shrinking, creating a gap that technology must help fill.
- Most age-tech fails because it lacks staying power; seniors disengage when tools feel like nagging reminders without a relationship.
- Aileen’s differentiator is relationship-driven AI delivered through simple phone calls—building familiarity first, then enabling reminders and support to stick.